DISCO Auto Review is an industry-leading document review platform used by litigation, compliance, and ediscovery teams. Auto Review is powered by DISCO’s cutting-edge legal AI, resulting in faster turnaround, lower costs, and increased precision during legal document review.

DISCO Auto Review streamlines the first-pass document review process using generative AI. It begins with your legal team, often in partnership with DISCO's AI experts, creating specific tag descriptions that define the criteria for document relevance, similar to a traditional review protocol.
‍
Once these tag instructions are set, Auto Review's underlying large language model (LLM) technology evaluates your document set against these criteria. The generative AI then suggests which documents are relevant to which tags.

Crucially, for each suggestion, Auto Review provides a detailed explanation, offering transparency into its reasoning. This allows your team to quickly understand the AI's decisions, efficiently conduct quality control, and advance to strategic case work much faster than with manual review methods.

In DISCO Auto Review, tags are defined by custom descriptions you create, outlining what makes a document relevant for a particular issue or category. These descriptions serve as instructions for the AI. You can test and iteratively refine these tag descriptions on sample documents, with support from DISCO's AI consulting team, to ensure the AI understands the nuances of your specific matter and achieves high recall and precision rates.

Tag suggestions in DISCO Auto Review are the result of its generative AI analyzing documents based on your custom tag descriptions. For each document, the AI proposes which tags should apply. Importantly, Auto Review doesn't just provide a suggestion; it also offers detailed explanations for each tagging decision. This transparency allows your team to understand the AI’s rationale, confidently verify its accuracy, and efficiently manage the quality control process.

DISCO Auto Review can process documents at an precedented speed, reviewing up to 32,000 documents per hour.

DISCO Auto Review generally delivers faster, more consistent, and more accurate results than traditional human review. It can achieve recall and precision rates 10-20% higher than published rates for human reviewers.

The Large Language Models (LLMs) used in DISCO Auto Review are prohibited from training on or retaining any customer data after processing. Furthermore, all LLM vendors must meet DISCO's stringent SOC Type 2 or ISO 27001 product certifications, ensuring your data's confidentiality and integrity. DISCO also maintains measures for GDPR compliance.